مجله پژوهش های مکانیک ماشین های کشاورزی
سال نهم شماره 2 (پاییز و زمستان 1399)

Nowadays, application of machine vision techniques had been extensively used in agriculture and particularly in food industries. This system can be used in quality separation, especially for valuable products such as walnut. The high nutritional value of walnuts has caused the crop to be widely processed in many processed foods. The texture is one of the most important features of agricultural crops which has been widely applied in food industry for quality evaluation. The texture of images reflects changes in pixel intensity values, which may include information from the geometric structure of objects. In this study, the possibility of walnut separating in three categories, based on quality, including: light-intact , dark-intact and damaged kernel using image processing and color systems such as RGB, HSV and L*a*b* on Kaghazi varieties was investigated. The machine vision system includes a lighting box, a camera (model SC-W12 SONY with resolution of 4 mega pixels), a computer and MATLAB software. So that, all samples are captured in RGB color system, then using transfer functions the other color systems components were calculated. Also, separation of intact samples from non-intact was evaluated using statistical analyzing on color space of RGB, L*a*b* and HSV. The results showed that in the RGB color space using of components of R (redness intensity) and G (green color intensity), and in the HSV color space based on component H and V, separation of healthy samples was possible. The success of this method was 92%. Also, in the L*a*b* color space, components of L* and b* be able to clear healthy samples from the other two categories. The success of this method was 90%. However, separating dark-intact samples from damaged samples were not possible because of overlapping of colors area. In case of surface tissue indices, contrast and energy were able to separate intact samples from non-intact samples.

Investigating the properties of a rice single kernel is one of the methods used to determine the quality of grain bulk. In addition, one of the major problems in predicting the quality properties of rice is the high cost of conducting laboratory experiments in food industry. Therefore, the objective of this study was to predict some quality properties of white rice via the use of bending test in three varieties of white rice (Khazar, Dorfak and Kadus) at two levels of moisture content 9 and 14% w. b. In this study, the response surface methodology (RSM) (historical data design) was used to model and estimate quality properties (Amylose content (AC), Protein content (PC), Gelatinization Temperature (GT), Gel Consistency (GC), Minimum viscosity (Min. V), Peak Viscosity (PV), Final Viscosity (FV), Breakdown viscosity (BD.V), Setback Viscosity (SB.V), Peak Time (PT) and Pasting temperature (Pasting. T)) of white rice through analyzing independent variables (strength value of rice single kernel in deformation of 0.05 mm, maximum rupture force, maximum deformation, rupture energy, tangent modulus at deformation of 0.05 mm, maximum tangent modulus, secant modulus at deformation of 0.05 mm and maximum secant modulus). The results of ANOVA, showed that all models are significant at probably levels of (P < 0.05) and (P < 0.01). According to the results, the most appropriate model for prediction of the response variables was the linear-Interaction (2FL) model whit the coefficient of determination (R2 ≥ 0.95). In addition, Optimum conditions of independent variables were determined based on the maximum goal or the minimization of response variables. The desirability degree of the optimization was (R2 ≥ 0.82). The results indicated that the response surface methodology can be quite useful in optimizing the models developed for predicting the quality properties of white rice from the strength properties of a single kernel.

The quail egg is a without cholesterol food and riches of different vitamins. This food expunges the radioactive particle and the carcinogen of human’s body. For mechanizing its preparation process, the quality parameters should be modeled with computer techniques. Therefore in this research, the image processing technique was used to determine the physical properties of quail egg, then egg volume was calculated by slicing each egg and rotating the slices around its main axis. Furthermore to determine the egg weight, a regression multi-variable model and a neural network model were developed. The results of suggested methods and models in this research were compared with real values and previous methods (Hoyt model). Finally, the influence of nutrition, ventilation and female quail age on quality parameters of eggs was determined. By comparing the real volume of eggs with the Hoyt model, accuracy was 97% and with the slices-rotation method, accuracy was 100% which well shows the importance of slices-rotation method. By comparing the real weight of eggs with the Hoyt model, accuracy was 83%, with the regression multi-variable model, accuracy was 90%, and with the neural network model, accuracy was 95%. The neural network model not only presented the egg weight with high accuracy but also considered the effect of laying time which is very important for the egg weight estimation. After comparing the results, it was shown that the nutrition, ventilation and female quail age influence egg quality. The eggs of younger females were limited to spherical geometry that may be due to the mother's womb not being expanded. Furthermore, the female quails in topper cages which had better nutrition and ventilation showed more weighting and sizing. These results show that coupled the image processing and artificial neural network techniques can help very well to mechanize the preparation process of quail egg.

In this research, the adaptive neuro-fuzzy inference system was used for predicting the parameters related to traction of tractor included of drawbar power (DP), drive wheels slippage (S), traction efficiency (TE) and overall energy efficiency (OEE) in tractor- implement combination under the effect of independent variables included of tine type (subsoiler and paraplow), depth (30, 40 and 50 cm) and forward speed (1.8, 2.3, 2.9 and 3.5 km/h) during subsoiling operation. The field data were used to create the regression and ANFIS models for predicting the studied parameters and the results of them were compared together. The field results showed that all independent variables were effective on the studied parameters except TE. The increment of forward speed and depth resulted in increase of S, DP, OEE and decrease of TE. Moreover, with considering the studied parameters, the paraplow tine was more commodious than subsoiler tine. The results of ANFIS part showed that about S, DP, TE and OEE, the membership functions of Trimf, dsigmf, Primf and Gaussmf with the mean square error of 0.0159, 0.0231, 0.0212 and 0.0224 also correlation coefficient of 0.9996, 0.9999, 0.9985 and 0.9997 caused the best models to predict, respectively. ANFIS models had higher accuracy than regression models and it could be calculated the model outlet for a special inlet using ANFIS outlet surfaces.

The use of digital image processing (DIP) has made it possible to exactly locate fruits to survey the performance. Moreover, advancements in the use of sensors and the robot technology have provided an opportunity to develop mechatronic crop harvesting systems as agricultural tools to facilitate gardening and harvesting mechanization. Walnuts are among crops that require proper and science-based management when harvesting. Traditional walnut harvesting has a high cost and high risk for Harvesting workers. In addition to the high cost and relatively low performance of walnut shakers, tall and large-trunk trees and irregular and rolling gardens make it difficult to widely employ mechanized harvesting technologies and machinery. One of the precision agricultural tools is crop yield mapping for local management and higher crop production. Product performance increases with local management based on understanding the causes of changes in different environments. Hence, to evaluate performance and build a harvesting machine, a DIP algorithm is proposed to detect and locate walnuts on trees in a natural light condition. The proposed algorithm functions based on image texture characteristics, light and shade density, and edge detection. The use of texture characteristics in combination with the light and shade density as an innovative method has worked well in conditions where fruits and leaves have a common color spectrum. In this method, fruit detection was independent of its shape to minimize the effect of fruits appearance obstruction on the algorithm's performance. The results indicated that the proposed method was able to detect walnuts in images taken from trees with a detection rate of 90.8%. The fruits that were next to each other were also distinguished at 93% resolution. Walnut detection-inhibiting factors were image darkness-induced errors with 2.7%, low resolution with 2.7%, improper scale with 1.8%, distinguish the tandem walnuts with 1.8%, and walnut appearance obstruction with 0.9%

Water scarcity challenge and the high rate of population growth means food production needs to be increased, therefore using proper management practices which can improve water use efficiency (WUE), increase grain yield and reduce carbon dioxide emission from soil, seem to be necessary. In this research, effects of tillage systems and planting methods on water use efficiency, carbon dioxide emission from soil and wheat yield was studied at Agricultural Science and Natural Resources University of Khuzestan located in North of Ahwaz county in 2018-19. The study was conducted using a split complete randomized block design with 3 replications. Tillage at two levels including conventional tillage system (moldboard plow followed by offset disc) and minimum tillage system (combination tillage implement) as main factor and planting methods including flat and raised bed planting methods considered as sub factor. The results indicated that tillage systems and planting methods had a significant effect on water use efficiency at 5% and 1% levels, respectively. The results, also, showed that interaction of tillage systems and planting methods had a significant effect on water use efficiency at 5% level. The results related to yield showed that planting methods had significant effect on wheat yield, while effect of tillage systems and their interaction on wheat yield was not statistically significant. Minimum tillage significantly reduced carbon dioxide emission from soil compared to conventional tillage at 5% level. Although conventional tillage had a lower soil cone index at the beginning of wheat cropping season compared to minimum tillage, over time, soil cone index was approximately the same for both tillage levels. Considering the interaction effect, minimum tillage-raised bed planting treatment with a grain yield of 5426 kg/ha, water use efficiency of 1.13 kg/m3 and carbon dioxide emission of 13.33 kg. ha-1.day-1 was suggested as the best system for the studied region.

The estimation of ripeness is a significant section of quality estimation since maturity at harvest can affect on sensory and storage properties of fruits. A possible tactic for defining the grade of ipeness, sensing the aromatic volatiles released by fruit using e-nose. For detection the five ripeness grades of strawberry, the e-nose machine was designed and fabricated. Organic healthy samples were collected and divided into five ripeness grades (RG1 = totally ripe, RG2 = close to ripeness, RG3 = intermediate to ripeness, RG4 = close to unripe and RG5= unripe), according to the criteria used by expert growers (based on physical size and appearance as well as estimated maturity stages) during June2019. organic and inorganic sour cherries have different response patterns. This indicates that their aromatic compounds are different. It is clear that different degrees of maturity have different response patterns. This indicates that their aromatic compounds are different. Generally, in organic strawberry MQ3, TGS2602 and TGS822 sensors had the highest response and role in detecting of Ripeness Grades of Strawberry. Artificial neural networks (ANN), principal components analysis (PCA) and linear discriminant analysis (LDA) were applied for pattern recognition of array sensors. The best structure (10–11-5) can classify the samples in five classes in ANN analysis with precision of 98.3 for strawberry. Also, PCA analysis characterized 99% variance in the strawberry. The LDA distinguished the strawberry ripening grades well with just a little overlap between the RG4 and RG5. The accuracy of the analysis was 93.3%. Each three methods can be detected RGs, but PCA with correct classification percentage 99%, is the best method. According to the study, it can be expressed that the e-nose is a suitable instrument for detecting RGs of strawberry and can be used with less time and cost to determine the appropriate harvest time and RGs of strawberry.

A numerical model of wet soil-paddlewheel interaction was developed and validated against the experiments. The use of numerical computational methods to simulation of the wet soil-paddlewheel interaction reduces the experimental, design and development costs. In this study, SolidWorks software was used to design the paddlewheel and Abacus software was used for analysis. Coupled Eulerian-Lagrangian (CEL) method was applied for the analyses in which the soil and paddlewheel were introduced by Eulerian and Lagrangian formulation, respectively. Soil was modelled as an elastic- perfectly plastic material with Drucker-Prager yield function. The results showed that the wheel sinkage and slip increased by increasing drawbar pull. The stress inside the soil was shown to be increased due to the drawbar pull. Satisfactory agreement was found between the finite element model and experiments for slip; sinkage as well as rolling resistance. The finite element method was shown to be a confident method for wet soil-paddlewheel interaction and can be applied for effective parameter optimization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

One of the most important issues in progress and development of agricultural machinery is indebted to being automated. The on-line measurement of seed and granular fertilizer level in drills hopper could be employed in automation of agricultural machinery and equipment. In present study, a near infrared distance-meter sensor was adapted and tested in both laboratory and field conditions. The near infrared distance-meter sensor was first calibrated on three surfaces of wood, iron and soil within 100- 650 mm range in 50 mm increments. Having installed the sensor beneath drill hopper’s door, the static and dynamic evaluation was conducted for wheat grain within 100-450 mm and for triple supper phosphate within 100-250 mm range both in 10 mm increments. The laboratory results revealed that there was a nonlinear correlation with determination coefficient of 0.99 between sensor output and desired distance while it was used on three tested surfaces. Laboratory tests revealed that there was a minimum 100 mm distance limitation in installing the sensor to approach perfection. Calibration results showed that there was a nonlinear correlation with determination coefficient of 0.996 and 0.992 between the sensor output and the distance of wheat seed while it was tested in both static and dynamic conditions. Calibration results also showed that there was a nonlinear correlation with determination coefficient of 0.995 and 0.992 between the sensor output and the distance of fertilizer while it was tested in both static and dynamic conditions. In validation stage, determination coefficients of 0.983 and 0.961 were obtained for establishment of a linear regression between directed measured distance of wheat grain surface and corresponding distance detected by the sensor in static and dynamic conditions, respectively. Also, in validation stage, determination coefficients of 0.995 and 0.961 were obtained for establishment of a linear regression between directed measured distance of fertilizer surface and corresponding distance detected by the sensor in static and dynamic conditions, respectively. Therefore, the near infrared distance-meter sensor was considerably precise in detection of the granular materials’ level in drill hopper and could be applicable for automation of agricultural machinery and implements.

Agricultural production systems seriously changed in the world because of more dependent on chemical fertilizers and pesticides and hybrid seeds. As a result, Significant changes in the energy consumption in agriculture will be more focused on fossil fuels. Today, in addition to energy evaluation in agriculture, increasing greenhouse gas emission lead to growing considerable interest to climate change and global warming phenomena. The aim of this study was to determine the energy indices and greenhouse gas emissions from sugar beet production units in Chaharmahal and Bakhtiari province. In addition, in this study crop yield was predicted and modeled using artificial intelligence methods, such as artificial neural networks (ANN) and adaptive neuro-fuzzy inference system (ANFIS). Required data was collected through questionnaires and face-to-face interviews with farm owners and agricultural authorities. Simapro software was utilized to evaluate the environmental impacts. The Energy Ratio (ER), Energy Productivity (EP), Specific Energy (SE) and Net Energy Gain (NEG) were 20.99, 1.24 kg/MJ, 0.8 MJ/kg and 75244.93 MJ/ha, respectively. The total energy consumption and output energy (yield) for the production of sugar beet was obtained 37640.464 and 790090 MJ/ha respectively. The total amount of greenhouse gases emissions was estimated 1556.859 kg CO2 equivalent per hectare, that the largest share of total GHG emissions was owned to nitrogen fertilizers (40.22%), fuel (31.66%) and electricity (21.76%), respectively. Both ANN and ANFIS models significantly predict sugar beet product performance, but ANN have better and more accurate results. The correlation coefficient for prediction of sugar beet yield using ANN was 1, while the value of ANFIS was 0.9991. However, the ANN model had a lower error coefficient of 0.00003 compared to the ANNIS model. It is understood that by substituting optimal amounts of inputs for agricultural production, the negative effects on the environment (GHG emissions) will be reduced.

One of the methods of assessing the performance of seed drills may be to compare the performance with the crop population growth. Pixels of crop emergence zone appear to have similar characteristics concerning image parameter variations between soil and crop. The use of deep learning methods based on convolution neural networks to map regions of interest in the image seems appropriate. In this regard, a total of 2720 images of early-growth cereals were obtained from a field. 212 images with different backgrounds were selected and annotated to feed and train a neural network model. Raw images were defined as inputs and maps of manually marked growth points as network outputs. In order to calculate the network cost, the predicted output of the network was compared with the pre-marked pixel map. Prediction errors were then back-propagated and the network parameters updated. Examination of the initial network output showed that the trained network had responded incorrectly to plant tips, weeds and plant remains as plant growth points. To overcome these errors and improve network performance, a penalty function was defined for the mistaken predicted points. The network was trained with three penalty rates and evaluated with nine Soft Max thresholds. According to the network output, images were arranged in terms of plant density. In order to evaluate the model in different ranges, images from each particular range were selected at random. These images were fed to the model and their outputs compared with the truth. For the ranges where approximately 94% of the total field images existed, the average harmonic accuracy of the precision index and the recall index was estimated to be over 80%, indicating good model performance. The results showed that the model can provide acceptable feedback on sowing performance and improve farm management and efficiency in the next steps.

There are several ways to control noise pollution. Due to the spread of acoustic science, the use of impedance tubes and the use of sound absorbers are of particular importance. The simulation of several adsorbents is discussed in this paper. Fabric absorbers with small holes, Cavity-coated absorbers and… There are examples of these absorbers. Various factors such as material, step and frequency affect this absorber. In this study, the changes of sound level under material, step and frequency factors were studied, analyzed and compared in two experimental and simulation methods. Mainly simulation for different types of absorbers allows for better comparison and selection of the best absorber for sound control. In both experimental and simulation methods, the least changes in sound pressure level in MDF were observed. This value is reported for the experimental method of 70.3 dB and for the simulation of 76.5 dB. The smallest changes in sound level were reported in both methods at 128 Hz. The sound level changes for both methods were the same and from step one to step two sound pressure level changes increased. In laboratory method, one step of sound pressure level changes is 74.4 dB and in simulation method these changes are reported 80.4 dB. In step two, in both laboratory and simulation methods, the sound pressure level changes were 76.6 and 82.4 dB, respectively. Comparing the mean sound level changes for both ionolite and mineral wool, it is observed that in ionolite, the rate of these changes has increased from 25.9 dB to 27.8 dBfrom one to two steps, while in Mineral wool these changes have been reported in both steps with a constant value of 22.4 dB. therefore, the ionolite material can be introduced as the optimal absorber in terms of sound level changes in a wider range.